Many real estate investment practices ñ and much real estate analysis and research ñ have their starting point in categories: residential, commercial, office, industrial; retail and the non-category category, mixed use. Categories are social conventions with scarce ontological status. Thus, despite the development of increasingly sophisticated analytic tools, the epistemological foundations of large parts of real estate thought remain in a pre-scientific state. The result is a host of problems involving what appears to be rapid obsolescence in persistent office vacancies, failed retail centers and other problems usually attributed to market conditions and typically seen from the bottom up as part of business as usual in real estate. This paper proposes that the appropriate starting point for real estate thought be the configurational model rather than the category. It begins with examples of how categories are casually used in real estate research, reviews the problem of asset mispricing as an example of over-reliance on categories and examines the category issue from two points of view ñ folk thinking (everyday as opposed to scientific discourse) and evolutionary processes (things change and get to the edge of and fall out of the category while the names of them donít). It follows with a discussion of models-of and models-for as a basis for the modeling process and concludes with an application.